Press Release: Immediate Release
San
Jose, CA
Sightech
Vision Systems, Inc.
408
2823770
Attn: Art Gaffin or Tom Seitzler
Subject: Fluid flow contaminants inspection
San
Jose, CA September 14, 2005
The
Sightech PC Eyebot was recently applied successfully to an application that
demanded the finding of bubbles in a container of water. This seemingly simple application has important
implications for liquid inspections of many varieties.
The
requirement was to identify bubbles of any type, size, location, shape or
orientation within the segment of the camera’s working area. A simple 640 x 480 camera was used. There were all types of reflections on the
water, and a variety of waves in various directions and amplitudes from the
breaking bubbles. The PC Eyebot was
required to “LEARN” (train) on the bubbles, then “FORGET” (ignore) normal
conditions of the water.
The
three attached images show the process of training, forgetting, and then
inspecting. These images, additional
explanation, and a tutorial on setting up an application on the PC Eyebot can
be viewed at http://www.sightech.com/applications.html. The camera was placed; “emphasis” and
feature characteristics were selected to clearly define objects of
interest. The bubbles showed
clearly. Next the “FORGET” mode was
selected and the water, without the bubbles, was forgotten by the Eyebot
inspection processor. When the Eyebot
was placed in “RECOGNIZE” mode, bubbles were recognized and deemed “present”
where flat water, or water without the bubbles resulted in an “absent”
condition. The user can determine how
many bubbles are needed to trigger a “present” judgment.
The
random characteristics of the bubbles made this an especially difficult
application for any other type of
inspection system. The glare from the
water, and the continuously changing background, as well as variable lighting
added to the challenge.
Given
sensors with proper sensitivities, many contaminants can be detected either on
or in fluid flows of this type.
This
bubble detection is another example that is leading the market to agree with
the complimentary judgment: “If you or
I can recognize it, so can the PC Eyebot”.